Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror

Comment: Re:Write-only code. (Score 1) 757

by grahamwest (#49228595) Attached to: Was Linus Torvalds Right About C++ Being So Wrong?

I'm not a Linux programmer so I may be out of date on this, but there isn't or wasn't a single C++ ABI on Linux between the various compilers. If the kernel used C++ for those interfaces it would potentially require that the kernal and all kernel modules were compiled with the same toolchain. Rolling their own implementation means the ABI is compatible across all the different compilers and compiler version with a side benefit of being able to write kernel modules in languages other than C/C++.

Comment: Re:This sounds silly ... (Score 1) 411

by grahamwest (#49033287) Attached to: Your Java Code Is Mostly Fluff, New Research Finds

I read the paper and their premise seems to be that MINSETs can be browsable/searchable and good enough to let programmers figure out whether a given function is worth investigating further. Basically they're a better replacement for text search and class browsing.

I'm skeptical about that, especially looking at their examples, but I can't dismiss it outright. I think it might interact in a favourable way with metaprogramming techniques, whether C++ templates or Lisp macros, but that's just speculation.

Comment: Re:wow (Score 1) 571

by grahamwest (#48152289) Attached to: Lockheed Claims Breakthrough On Fusion Energy Project

Can you explain in more detail? It wasn't clear to me how this problem was handled. I did a little research and learned that the fast neutrons cause neutron activation, creating often long-lived radioactive isotopes of what they hit - which will generally be the reactor containment walls.

He did mention breeding tritium via lithium, so is the idea to plate the inner walls of the reactor with lithium? In that case, does the amount of tritium generated balance with the amount consumed? Or, does that just naturally reach equilibrium?

In any case, I think all of this is only alluded to in the video. If you have more insight, I think it would be useful to share it.

Comment: Plain old DOS (Score 1) 334

Build them a small form factor PC. Don't connect the USB ports and don't provide an optical drive. Install FreeDOS, a basic dialup TCP/IP stack, an email client and a web browser. I'd also set it up to decompress and restore everything from a read-only partition every single time they boot, ideally before processing config files. If there isn't already a way to do this, FreeDOS does have source available, so you could modify that.

Notes on DOS email and web clients:
http://www.tokyopc.org/newslet...
http://www.compmiscellanea.com...

Comment: Re:What about the lunch ladies? (Score 1) 282

They cover that in the full memo:

Q: Why do some supplier employees not take breaks when others do?

A: There are some business functions and processes that have been fully outsourced (Outsourcing), such as cafeteria services, landscaping and call centers. These Outsourcing engagements are limited, require a certain set of criteria be met and must go through a rigorous approval process.

Comment: Re:A few options (Score 1) 266

by grahamwest (#46608301) Attached to: Ask Slashdot: How To Handle Unfixed Linux Accessibility Bugs?

I just tested and OS X treats the clicks and key presses the same way when sticky keys is enabled. Hit the modifier, the next click or key press is modified. Hit the modifier twice and all clicks and key presses and modified until you hit the modifier again to unlock.

Seems very much the logical way to do it.

Comment: Re:was a fix to make follow the specification (Score 1) 266

by grahamwest (#46608267) Attached to: Ask Slashdot: How To Handle Unfixed Linux Accessibility Bugs?

The spec in question - http://www.x.org/docs/XKB/XKBp... as Peter references in the bug comments - discusses StickyKeys (4.4 on page 9) and strongly implies modifiers only unlatch on key presses; mouse buttons are not mentioned. His change made the code match this reading of the spec. I have a hard time believing that's what the spec writers intended, but if so then KDE's lock checkbox really isn't supposed to do anything.

Comment: Re:Did Fluke request this? (Score 2) 653

by grahamwest (#46526621) Attached to: $30K Worth of Multimeters Must Be Destroyed Because They're Yellow

This was my reaction as well. I looked at the trademark registration, which has a picture of the Fluke, then at Sparkfun's site. So, fair enough. However, I google image searched 'multimeter' and there are lots of multimeters in that same shade of yellow, of all sorts of brands. I had no idea yellow "meant" Fluke, personally. I think there's a valid case that this trademark has become diluted and generic. Whether all those others are licensed uses or not, if there's no scope for customer confusion of brand, it's no longer a valid trademark.

Comment: VERY misleading terminology (Score 5, Informative) 676

by grahamwest (#46457753) Attached to: 70% of U.S. Government Spending Is Writing Checks To Individuals

"Individual" in this case does NOT mean "person".

If you download the spreadsheet you can see that they classify total spending as either "direct" or "grants", of which the vast bulk is "direct". Everything that is not a grant must be being paid to an entity of some kind, whether an actual person, a company, a non-profit or something. You can verify this is the total Federal spending using the Monthly Treasury Statements at https://www.fms.treas.gov/mts/... - I recommend the PDF versions.

As for the percentage going to veterans, I expect the number of veterans isn't growing very much, whereas the Federal budget is. So a constant amount in a larger total is going to be a smaller fraction.

Bottom line, this article is FUD and should not be taken serious by anyone.

Comment: Re:So does anybody... (Score 4, Insightful) 111

by grahamwest (#44838711) Attached to: Cisco Can't Shield Customers From Patent Suits, Court Rules

The patents are about automatic failover when network nodes or spans break. The earlier patents are about having spare nodes and spans and deciding which to use when some part of the network fails (eg. having a node which broadcasts "who can help?" and available nodes broadcast back "i can help!" and a single node decides which available node to use). The later patents are about turning on and off routes between nodes to reconfigure the network, usually into some sort of mesh network.

I'm not a network engineer so it's hard for me to judge, but the earlier patents seem trivial to me especially since they're from the late 1990s. The latter patents might have some merit - the idea of changing the network to a mesh is interesting, but my gut feeling is they're mostly solutions that any decent engineer would come up with after a bit of head scratching.

Comment: Re:federal overreach, as usual (Score 1) 356

by grahamwest (#44787989) Attached to: Indiana Man Gets 8 Months For Teaching How To Beat Polygraph Tests

If I come to you and ask you to sell me some dynamite so I can rob a bank with it, it doesn't matter whether I actually rob the bank or not. By agreeing to sell me the dynamite you become part of a criminal conspiracy. There's no duty on you to tell the cops about me, but there is the duty to say, "Sorry dude, I can't help you out with that".

Real Programmers don't write in FORTRAN. FORTRAN is for pipe stress freaks and crystallography weenies. FORTRAN is for wimp engineers who wear white socks.

Working...